5 Definite integral

advertisement
5
Definite integral
5.1
Definition of definite integral
There are many ways of formally defining an integral, not all of which are
equivalent. The differences exist mostly to deal with differing special cases
which may not be integrable under other definitions, but also occasionally
for pedagogical reasons. The most commonly used definition of integral is
Riemann integral.
Assume that the function y = f (x) is defined on the closed interval [a; b].
Let
a = x0 < x1 < x2 < . . . < xk−1 < xk < . . . < xn = b
be an arbitrary (randomly selected) partition of the interval [a; b], which
divides the interval into n subintervals
[xk−1 ; xk ], where k = 1, 2, . . . , n
Denote by ∆xk = xk − xk−1 the length of the kth subinterval.
Further we choose on any of these subintervals an arbitrary point
ξk ∈ [xk−1 ; xk ], k = 1, 2, . . . , n
and multiply the value of the function at ξk by the length of the kth subinterval:
f (ξk )∆xk , k = 1, 2, . . . , n.
Assuming f (x) ≥ 0 on [a; b], this product is the area of the rectangle with
height equal to the function value at the distinguished point ξk of the kth
sub-interval, and width the same as the sub-interval width ∆xk .
y
y = f (x)
f (ξk )
a
xk−1 ξk xk
b
x
Figure 5.1: the area of the rectangle f (ξk )∆xk
1
If we add these products, we obtain the sum
sn =
n
∑
f (ξk )∆xk
k=1
which is called the integral sum of the function f (x) over the interval [a; b].
The dividing points x1 , x2 , . . . are chosen arbitrarily. Hence, the subintervals have the different length ∆xk , k = 1, 2, . . . , n. Let λ denote the
greatest length of these subintervals
λ = max ∆xk
1≤k≤n
Definition 1. If the limit
lim sn
λ→0
exists and does not depend on the partition of the interval [a; b] and does not
depend on the choice of the points ξk ∈ [xk−1 ; xk ], then this limit is called
the definite integral of the function f (x) from a to b and denoted
∫b
f (x)dx
a
∫
The symbol is the integral sign, a is called lower limit, b is called upper
limit, f (x) is called integrand and dx is called the differential of the argument,
x is called the variable of integration.
Definition 2. If the conditions of the definition 1 are satisfied then f (x)
is called the integrable function on [a; b].
There holds the following theorem.
Theorem 1. If the function f (x) is continuous on [a; b] then it is integrable on [a; b].
Remark. If the function is discontinuous on [a; b] then it may be integrable or may be not integrable on [a; b].
By definition
∫b
n
∑
f (x)dx = lim
f (ξk )∆xk
λ→0
a
k=1
If f (x) ≥ 0 on the interval [a; b], then the products f (ξk )∆xk in the
integral sum are the ares of the rectangles with height f (ξk ) and width ∆xk .
Hence, the integral sum represents approximately the area under the graph
of the function y = f (x), i.e. area of the domain in xy-plane bounded by
2
y
y = f (x)
a x1 x2
xk−1xk
xn−1b
x
Figure 5.2: The integral sum
the graph of the function y = f (x), x-axis and two vertical lines x = a and
x = b.
If λ → 0, then the length of any subinterval gets shorter and, to cover the
given interval [a; b], the number of those subintervals has to increase. The
integral sum - the sum of the areas of the rectangles - is getting closer to the
area under the graph of the function y = f (x).
Therefore, if f (x) ≥ 0 on the interval [a; b] then the definite integral is
the area under the graph of the function y = f (x).
y
y = f (x)
∫
b
f (x)dx
a
a
b
x
Figure 5.3: The area under the graph of the function y = f (x)
3
5.2
Properties of definite integral
Property 1 The definite integral of the sum of two functions is equal to
the sum of the definite integrals of these functions:
∫b
∫b
[f (x) + g(x)]dx =
a
∫b
f (x)dx +
g(x)dx.
a
(2.1)
a
Proof By the definition of the definite integral
∫b
I=
[f (x) + g(x)]dx = lim
λ→0
a
n
∑
[f (ξk ) + g(ξk )]∆xk
k=1
Removing the square brackets under the sum and taking into account that
the sum does not depend on the order of summation, we obtain
]
[ n
n
∑
∑
g(ξk )∆xk
f (ξk )∆xk +
I = lim
λ→0
k=1
k=1
The limit of the sum equals to the sum of the limits, i.e.
I = lim
λ→0
n
∑
f (ξk )∆xk + lim
λ→0
k=1
n
∑
g(ξk )∆xk
k=1
By the definition of the definite integral these limits are the definite integrals
on the right side of (2.1).
Property 2 The constant coefficient c can be factored out:
∫b
∫b
cf (x)dx = c
a
f (x)dx
a
The proof is similar to the proof of the first property.
Conclusion 1. The definite integral of the difference of two functions
equals to the difference of the definite integrals of those functions:
∫b
∫b
[f (x) − g(x)]dx =
a
∫b
f (x)dx −
a
g(x)dx
a
Proof It follows from the first and the second property. Writing f (x) −
g(x) = f (x) + (−1)g(x) gives
∫b
∫b
[f (x) − g(x)]dx =
a
∫b
[f (x) + (−1)g(x)]dx =
a
f (x)dx + (−1)
a
4
∫b
g(x)dx
a
which is we wanted to prove.
Property 3 If f (x) ≥ 0 for any x ∈ [a; b], then
∫b
f (x)dx ≥ 0
a
Proof. If f (x) ≥ 0 on [a; b], then f (x) ≥ 0 on any subinterval [xk−1 ; xk ],
k = 1, 2, . . . , n. Thus, for ξk ∈ [xk−1 ; xk ] also f (ξk ) ≥ 0. Multiplying
the last inequality by the length of the kth subinterval gives f (ξk )∆xk ≥ 0,
k = 1, 2, . . . , n.
Adding n nonnegative quantities, we obtain nonnegative quantity
n
∑
f (ξk )∆xk ≥ 0
k=1
By the limit theorem the limit of the nonnegative quantity as λ → 0 is
nonnegative, which proves the property.
Conclusion 2. If f (x) ≤ g(x) for any x ∈ [a; b], then
∫b
∫b
f (x)dx ≤
g(x)dx
a
a
Proof. By assumption g(x) − f (x) ≥ 0. Hence, by the Property 3
∫b
[g(x) − f (x)]dx ≥ 0
a
By Conclusion 1
∫b
∫b
f (x)dx ≥ 0
g(x)dx −
a
a
which proves the statement.
Property 4. The absolute value of the definite integral of the function
f (x) is less than, or equal to, the definite integral of the absolute value of
this function:
b
∫b
∫
f (x)dx ≤ |f (x)|dx
a
a
5
Proof. Here we use the property of the absolute value of the sum |a + b| ≤
|a| + |b| for n addends. By the definition of the definite integral
b
∫
n
n
∑
∑
f (x)dx = lim
f (ξk )∆xk = lim f (ξk )∆xk ≤
λ→0
λ→0 k=1
k=1
a
≤ lim
λ→0
n
∑
|f (ξk )∆xk | = lim
λ→0
k=1
n
∑
∫b
|f (ξk )|∆xk =
k=1
|f (x)|dx
a
Property 5. If we change the limits of integration, then the sign of the
integral changes:
∫a
∫b
f (x)dx = − f (x)dx
a
b
∫a
f (x)dx, then the start point
Proof. If we define the definite integral
b
is b. If we assume that the definite integrals in this property exist, the limit
does not depend on the partition. So in both definitions we can use the same
partition. As well we can use in both definitions the same arbitrarily chosen
points ξ1 , ξ2 , . . . , ξk−1 , ξk , . . . , ξn , because the limit does not depend on
the choice of these points. If we move in direction from b to a, then the first
partition point is xn = b, next xn−1 , ..., xk , xk−1 , ..., x0 = a. The start point
of the kth subinterval xk and the endpoint xk−1 . By the definition of the
definite integral
∫a
f (x)dx = lim
λ→0
b
1
∑
f (ξk )(xk−1 − xk )
k=n
The integral sum in this definition is
1
∑
f (ξk )(xk−1 − xk ) =
k=n
1
∑
f (ξk )(−∆xk ) = −
k=n
1
∑
f (ξk )∆xk
k=n
and, because the sum does not depend on the order of addition,
=
1
∑
f (ξk )(xk−1 − xk ) = −
k=n
n
∑
k=1
6
f (ξk )∆xk
∫a
The limit of the left side of this equality as λ → 0 is
b
∫b
of the right side is −
f (x)dx and the limit
f (x)dx.
a
Conclusion 3. If the lower and upper limit of the definite integral are
equal, then the definite integral equals to zero:
∫a
f (x)dx = 0
a
Proof. Changing the limits of integration, we have by Property 5
∫a
∫a
f (x)dx = −
a
f (x)dx
a
or
∫a
f (x)dx = 0
2
a
which yields the assertion.
Property 6 (Additivity property of the definite integral).
∫b
∫c
f (x)dx =
a
∫b
f (x)dx +
a
f (x)dx
c
Proof. First we assume that c is in the interval [a; b], i.e. a < c < b.
Defining the integral on the left side of this equality, we choose an arbitrary
partition of the interval [a; b], so that the first partition point is c. The further
arbitrary partition of [a; b] produces an arbitrary partition of the intervals
[a; c] and [c; b]. Thus, the integral sum for the whole interval [a; b] can be
written as the sum of the two integral sums
∑
∑
∑
f (ξk )∆xk =
f (ξk )∆xk +
f (ξk )∆xk
[a;b]
[a;c]
[c;b]
If the greatest length of the subintervals of [a; b] λ → 0, then the greatest
lengths of the subintervals of [a; c] and [c; b] approach zero as well. Therefore,
taking the limits as λ → 0 on both sides completes the proof.
7
If c is outside the interval [a; b], suppose c > b > a, then
∫c
∫b
f (x)dx =
It follows
∫c
f (x)dx +
f (x)dx
a
a
b
∫b
∫c
∫c
f (x)dx −
f (x)dx =
a
a
f (x)dx
b
and changing the limits we have by Property 5
∫c
∫b
f (x)dx =
a
∫b
f (x)dx +
a
f (x)dx
c
In similar way we can prove that this property holds if c < a.
Property 7. If m is the least value of f (x) and M is the greatest value
of f (x) on the interval [a; b], then
∫b
f (x)dx ≤ M (b − a)
m(b − a) ≤
a
Proof. The proofs of these two inequalities are similar and we prove only
the right hand inequality.
As assumed, the greatest value of the function f (x) on [a; b] is M . Thus,
f (ξk ) ≤ M for any arbitrarily chosen ξk ∈ [xk−1 ; xk ] for each k = 1, 2, . . . , n.
Multiplying this inequality by ∆xk gives
f (ξk )∆xk ≤ M ∆xk
Adding these products, we obtain
n
∑
f (ξk )∆xk ≤
k=1
n
∑
M ∆xk =
k=1
= M (x1 − x0 + x2 − x1 + x3 − x2 + . . . + xn − xn−1 ) = M (b − a),
because x0 = a and xn = b.
There is constant on the right side of the inequality
n
∑
f (ξk )∆xk ≤ M (b − a)
k=1
8
and taking the limit on both sides of this inequality as λ → 0 gives the
assertion.
Property 8 (Mean value property of the definite integral). If
the function f (x) is continuous on [a; b], then there exists at least one point
ξ ∈ [a; b] such that
∫b
f (x)dx = f (ξ)(b − a)
a
Proof. The function continuous in the closed interval has the least value m
and the greatest value M on this interval, hence, there holds the Property 7.
Dividing the both sides of the both inequalities by the length of the interval
of integration b − a gives
1
m≤
b−a
Consequently,
1
b−a
∫b
f (x)dx ≤ M
a
∫b
f (x)dx
a
is between the least and the greatest value. The function continuous on [a; b]
has any value between the least and the greatest. Therefore, there exists at
least one point ξ ∈ [a; b], where the function obtains this value, that is
1
f (ξ) =
b−a
∫b
f (x)dx.
(2.2)
a
The multiplication of both sides of this equality by b − a completes the proof.
The value f (ξ) is called the mean value of the function f (x) on the interval
[a; b]. This value is computed by (2.2).
5.3
Computation of definite integral. Newton-Leibnitz
formula
Suppose f (x) is defined on [a; b]. Let us define on [a; b] the function of
the upper limit of the definite integral
∫x
Φ(x) =
f (t)dt
a
9
(3.3)
Theorem 1. If the function f (x) is continuous on [a; b], then Φ(x) is
differentiable on (a; b) and Φ′ (x) = f (x).
Proof. We use the definition of the derivative of Φ(x)
Φ(x + ∆x) − Φ(x)
∆x→0
∆x
Φ′ (x) = lim
By additivity property of the definite integral,
x+∆x
∫
Φ(x + ∆x) − Φ(x) =
∫x
f (t)dt −
a
∫x
x+∆x
∫
a
a
x+∆x
∫
∫x
f (t)dt −
f (t)dt +
x
f (t)dt =
f (t)dt =
a
f (t)dt.
x
As assumed, the function f (x) is continuous on [a; b]. Hence, by the mean
value property, there exists ξ ∈ [x; x + ∆x] such that
Φ(x + ∆x) − Φ(x) = f (ξ)(x + ∆x − x) = f (ξ)∆x
Consequently
Φ(x + ∆x) − Φ(x
= f (ξ)
∆x
In the definition of the derivative ∆x → 0. It follows that x + ∆x → x
and since ξ is a point between x and x + ∆x, then ξ → x also. Thus,
Φ′ (x) = lim f (ξ) = lim f (ξ)
∆x→0
ξ→x
and the third condition of continuity of f (x) gives Φ′ (x) = f (x), which is we
wanted to prove.
Remark. In some textbooks the Theorem 1 is referred as the Fundamental Theorem of Calculus.
By Theorem 1, the function Φ(x) is an antiderivative of f (x). If F (x)
is the known antiderivative of f (x) (by the table of integrals or by some
technique of integration), then by Corollary 1.2 of 4.1 the antiderivatives
Φ(x) and F (x) differ at most by a constant, i.e. Φ(x) = F (x)+C. According
to the definition of Φ(x)
∫x
F (x) + C =
f (t)dt.
a
10
(3.4)
Taking in this equality x = a, we obtain by Conclusion 3 of previous
subsection
∫a
F (a) + C = f (t)dt = 0
a
which yields C = −F (a). Substituting C to (3.4) gives
∫x
F (x) − F (a) =
f (t)dt
a
and taking in the last equality x = b, we obtain
∫b
F (b) − F (a) =
f (t)dt.
(3.5)
a
Consequently, the antiderivative familiar from the indefinite integral is the
appropriate tool to evaluate the definite integral. Now we take in (3.5) for
the variable of integration x again. To facilitate the computation we use the
notation
b
F (b) − F (a) = F (x)
a
Finally, we formulate the result obtained as a theorem.
Theorem 2. If the function f (x) is continuous on [a; b] and F (x) is the
antiderivative of f (x), then
∫b
b
f (x)dx = F (x) = F (b) − F (a),
a
a
The formula (3.6) is called Newton-Leibnitz formula.
Example 1. Evaluate
∫e
1
e
dx
= ln x = ln e − ln 1 = 1
x
1
Example 2. Evaluate
∫1
xdx
√
1 + x2
0
11
(3.6)
For the integration we use the equality d(1 + x2 ) = 2xdx and find
∫1
1
xdx
√
=
2
2
1+x
0
∫1
0
∫1
d(1 + x2 )
√
=
1 + x2
0
1
√
1 √
= · 2 1 + x2 = 2 − 1.
2
0
1
2xdx
√
=
2
2
1+x
Example 3. Compute the mean value of the function f (x) = x2 on [1; 3].
By the mean value formula (2.2), we find
1
3−1
5.4
∫3
3
(
)
1 27 1
13
1 x3 1
x dx =
=
−
=
=4
2 3 1 2 3
3
3
3
2
1
Change of variable in definite integral
The choice of the new variable depends on the function to be integrated.
These principals are familiar from the indefinite integral.
If we compute the definite integral, we are interested in its value, not in
the antiderivative of the initial function. This is because after the integration
by change of variable in the definite integral we don’t re-substitute the initial
variable. Instead of it we compute the limits of integration for the new
variable.
Changing the variable x = φ(t) in the definite integral
∫b
f (x)dx
a
we find dx = φ′ (t)dt. The equation φ(t) = a gives the lower limit for the
new variable t = α and the equation φ(t) = b gives the upper limit t = β.
The change of variable formula is
∫b
∫β
f (x)dx =
a
f [(φ(t)]φ′ (t)dt
α
∫2 √
Example 4. Compute I =
8 − x2 dx. To remove the irrationality we
0
√
√
change the variable x = 2 2 sin t. Then dx = 2 2 cos tdt and
√
√
√
√
8 − x2 = 8 − 8 sin2 t = 8 cos2 t = 2 2 cos t
12
We determine the limits for the new variable t. If x = 0,
√ then sin t = 0, it
√
π
2
follows t = 0. If x = 2, then 2 2 sin t = 2 or sin t =
, hence, t = .
2
4
Thus,
π
π
∫4 √
∫4
√
I =
2 2 cos t · 2 2 cos tdt = 8 cos2 tdt =
0
0
π
4
π
4
∫
= 4
π
∫
(1 + cos 2t)dt = 4
0
∫4
dt + 2
0
0
π
π
4
4
= 4t + 2 sin 2t = π + 2.
0
5.5
cos 2td(2t) =
0
Integration by parts
Let u(x) and v(x) be two differentiable function on [a; b]. The differential
of the product of these functions
d(uv) = [u(x)v(x)]′ = [u′ (x)v(x)+u(x)v ′ (x)]dx = u′ (x)v(x)dx+u(x)v ′ (x)dx = udv+vdu
By Conclusion 1.6 of subsection 4.1 the product uv is one of the antiderivatives of d(uv). Integration of the equality d(uv) = udv + vdu over
[a; b] gives
b ∫b
∫b
uv = udv + vdu
a
which yields
∫b
a
a
b ∫b
udv = uv − vdu
a
a
(5.7)
a
This is the formula of the integration by parts for the definite integral. The
choice of the function u and the differential dv is the same as in case of the
indefinite integral.
∫e
Example 5. Compute ln xdx
1
Here we choose u = ln x and dv = dx. Hence, du =
13
dx
and v = x. By
x
the integration by parts formula (5.7),
∫e
1
5.6
e ∫ e
e
dx
ln xdx = x ln x − x ·
= e − x = e − (e − 1) = 1
x
1
1
1
Improper integral over infinite interval
In this section we consider a couple of different kinds of integrals. Both
of these are integrals that are called improper integrals. In the first kind of
improper integrals one or both of the limits of integration are infinity.
Definition 1. Let the function f (x) be defined and continuous on the
infinite interval [a; ∞). If for any N ∈ [a; ∞) there exists the definite integral
∫N
∫N
f (x)dx and there exists the limit lim
f (x)dx, then this limit is called
N →∞
a
a
∫∞
f (x)dx.
the improper integral with the infinite upper limit and denoted
a
By Definition 1
∫N
∫∞
f (x)dx = lim
f (x)dx
N →∞
(6.8)
a
a
If the limit exists and is a finite number, then the improper integral is said
to be convergent. If the limit does not exist or the limit is infinite, then the
improper integral is said to be divergent.
Thus, to compute the improper integral, we first have to compute the
definite integral over [a; N ] and next find the limit of this result as N → ∞.
∫∞
dx
Example 6. Evaluate
.
1 + x2
0
By the formula (6.8),
∫∞
0
dx
= lim
1 + x2 N →∞
∫N
0
N
π
dx
= lim arctan x = lim (arctan N −arctan 0) =
2
N →∞
N →∞
1+x
2
0
So, this improper integral is convergent.
Definition 2.Let the function f (x) be defined and continuous on the
∫b
infinite integral (−∞; b]. If for any M ∈ (−∞; b] there exists
f (x)dx and
M
14
∫b
f (x)dx, then this limit is called the improper
there exists the limit lim
M →−∞
M
∫b
integral with the infinite lower limit and denoted
f (x)dx.
−∞
By Definition 2,
∫b
∫b
f (x)dx = lim
f (x)dx
M →−∞
−∞
(6.9)
M
The convergence and divergence of this improper integral are defined in
the same way as in the pervious case.
Definition 3. If the function f (x) is defined and continuous in (−∞; ∞),
then the improper integral over (−∞; ∞) is defined as
∫∞
∫c
f (x)dx =
−∞
∫∞
f (x)dx +
−∞
f (x)dx
c
where c is any finite real number.
If both of the improper integrals on the right side of this equality are
convergent, then this improper integral is said to be convergent. If at least
one of the improper integrals on the right side of this equality is divergent,
then this improper integral is said to be divergent.
Example 7. Let a > 0 and let us decide for which values of α the
∫∞
dx
improper integral
is convergent and for which values of α it is divergent.
xα
a
Denote this improper integral by I and find
∫∞
I=
dx
= lim
xα N →∞
a
∫N
dx
xα
a
If α ̸= 1, then
N
( −α+1
)
N
a−α+1
x−α+1 = lim
−
I = lim
N →∞
N →∞ −α + 1 −α + 1 −α + 1
a
If α > 1, then
(
lim
N →∞
1
1
−
α−1
(1 − α)N
(1 − α)aα−1
15
)
=
1
(α − 1)aα−1
that means, the improper integral is convergent. If α < 1, then
( 1−α
)
N
a1−α
lim
−
=∞
N →∞
(1 − α) 1 − α
i.e. the improper integral is divergent.
If α = 1, then
N
∫∞
dx
= lim ln x = lim (ln N − ln a) = ∞
N →∞
N →∞
x
a
a
thus, the improper integral is divergent again.
∫∞
dx
Consequently, the improper integral
is convergent, if α > 1 and
xα
a
divergent, if α ≤ 1.
In many cases we are rather interested in the convergence of the improper
integral than in the actual value of this integral. Moreover, sometimes an
improper integral is too difficult to evaluate, but we still need to know, is it
convergent or not. One technique is to compare it with a known integral.
The theorems below, called the comparison theorems, enable us to decide
whether the improper integral is convergent or divergent. We formulate these
theorems for the improper integral with infinite upper limit. These theorems
hold as well for the improper integrals with infinite lower limit and in case,
if both limits are infinite.
∫∞
φ(x)dx is
We assume, that we know whether the improper integral
a
convergent or divergent.
Theorem 3. Suppose that f (x) and φ(x) are two continuous on [a; ∞)
functions such that 0 ≤ f (x) ≤ φ(x) on this interval. Then the convergence
of the improper integral
∫∞
φ(x)dx
(6.10)
a
yields the convergence of the improper integral
∫∞
f (x)dx.
(6.11)
a
Suppose that f (x) and φ(x) are two continuous on [a; ∞) functions such
that 0 ≤ φ(x) ≤ f (x) on this interval. Then the divergence of the improper
integral (6.10) yields the divergence of the improper integral (6.11).
16
Theorem 4. Suppose that two continuous on [a; ∞) functions f (x) and
φ(x) are equivalent in the limiting process x → ∞. Then the convergence
of (6.10) yields the convergence of (6.11) and the divergence of (6.10) yields
the divergence of (6.11).
Definition 4. The improper integral (6.11) is called absolutely convergent, if the improper integral
∫∞
|f (x)|dx
a
is convergent.
Theorem 5. The absolute convergence of (6.11) yields the convergence
of this improper integral.
Example 8. Decide on the convergence or divergence of
∫∞
arctan xdx
1 + x2
1
π
In the half-interval [0; ∞) there holds arctan x ≤ . By Example 6, the
2
∫∞
dx
improper integral
is convergent. Applying Theorem 3 for f (x) =
1 + x2
0
arctan x
π
1
and φ(x) = ·
, we conclude that the given improper integral
2
1+x
2 1 + x2
is convergent.
∫∞
dx
Example 9. Decide on the convergence or divergence of
.
x−1
2
1
1
In the limiting process x → ∞, the functions f (x) =
and φ(x) =
x−1
x
are equivalent because
1
x−1
lim
x→∞ 1
x
x
=1
x→∞ x − 1
= lim
∫∞
By Example 7 the improper integral
dx
is divergent. Thus, by Theorem
x
2
4, the given improper integral is also divergent.
∫∞
Example 10. Decide on the convergence or divergence of
1
17
sin xdx
.
x2
sin x 1
For any x ∈ R there holds 2 ≤ 2 . By Example 7 the improper intex
x
∫∞
dx
gral
is convergent. By Theorem 3 this improper integral is absolutely
x2
2
convergent and by Theorem 5 it is convergent.
5.7
Improper integrals of unbounded functions
Suppose that the function f (x) is unbounded in a neighborhood of the
right endpoint b of the interval [a; b].
∫b−ε
Definition 5. If for any ε > 0 there exists the definite integral
f (x)dx
a
∫b−ε
and there exists the limit lim
f (x)dx, then this limit is called the improper
ε→0
a
∫b
integral of the unbounded function at the upper limit and denoted
f (x)dx.
a
By Definition 5 we evaluate the improper integral of the unbounded function in the neighborhood of the upper limit b, using the formula
∫b
∫b−ε
f (x)dx = lim
f (x)dx
(7.12)
ε→0
a
a
Improper integrals are often written symbolically just like standard definite integrals.
Suppose that the function f (x) is unbounded in a neighborhood of the
left endpoint a of the interval [a; b].
∫b
f (x)dx
Definition 6. If for any ε > 0 there exists the definite integral
a+ε
∫b
and there exists the limit lim
ε→0
a+ε
f (x)dx, then this limit is called the improper
∫b
integral of the unbounded function at the lower limit and denoted
f (x)dx.
a
18
By Definition 6 the improper integral of the unbounded function at the
lower limit a we evaluate by the formula
∫b
∫b
f (x)dx = lim
ε→0
a+ε
a
f (x)dx
(7.13)
If the function f (x) is unbounded in some interior point c of [a; b], then
we use the additivity property on the integral and write
∫b
∫c
f (x)dx =
∫b
f (x)dx +
a
a
f (x)dx
c
and evaluate the first addend by (7.12) and the second addend by (7.13)
If the limits in (7.12) and (7.13) are finite, then the improper integral is
said to be convergent. If these limits either does not exist or are infinite,
then this improper integral is said to be divergent.
Definition 7. The improper integral of the unbounded function is said
to be absolutely convergent if the improper integral
∫b
|f (x)|dx
a
is convergent.
Example 11. Let us find how the convergence or divergence of
∫b
a
dx
(b − x)α
(7.14)
depends on the exponent α.
1
The integrand
is unbounded in the neighborhood of the upper
(b − x)α
limit b. By formula (7.12)
∫b
a
dx
= lim
(b − x)α ε→0
∫b−ε
a
dx
.
(b − x)α
Suppose α ̸= 1. Using the equality of the differentials d(b − x) = −dx, we
19
obtain
∫b−ε
lim
ε→0
a
b−ε
∫b−ε
(b − x)−α+1 −α
(b − x) d(b − x) = − lim
=
ε→0
−α + 1 a
]
]
[a −α+1
[
(b − a)−α+1
ε1−α
ε
(b − a)1−α
= − lim
−
= lim
−
.
ε→0 −α + 1
ε→0
−α + 1
1−α
1−α
dx
= − lim
ε→0
(b − x)α
If α > 1, then α − 1 > 0 and lim εα−1 = 0. Hence,
ε→0
ε1−α
1
lim
=∞
= lim
ε→0 1 − α
ε→0 (1 − α)εα−1
that means the improper integral is divergent.
If α < 1, then 1 − α > 0 and lim ε1−α = 0, thus,
ε→0
[
]
(b − a)1−α
ε1−α
(b − a)1−α
lim
−
=
,
ε→0
1−α
1−α
1−α
i.e. the improper integral is convergent.
If α = 1, then
∫b−ε
lim
ε→0
a
dx
= − lim
ε→0
(b − x)α
∫b−ε
a
b−ε
d(b − x)
= − lim ln |b − x|
=
ε→0
b−x
a
= lim (ln |b − a| − ln |ε|) = ∞,
ε→0
i.e. the improper integral is divergent.
Consequently, the improper integral (7.14) is convergent if α < 1ja and
divergent if α ≥ 1.
For the improper integrals of unbounded functions there hold the analogous theorems as for the improper integrals over unbounded intervals.
Theorem 3’. If the functions f (x) and φ(x) continuous in the halfinterval [a; b) satisfy the condition 0 ≤ f (x) ≤ φ(x) then the convergence of
the improper integral
∫b
φ(x)dx
(7.15)
a
yields the convergence of the improper integral
∫b
f (x)dx
a
20
(7.16)
Theorem 4’. If the functions f (x) and φ(x) continuous in the halfinterval [a; b) are equivalent in the limiting process x → b then the convergence of (7.15) yields the convergence of (7.16) and the divergence of (7.15)
yields the divergence of (7.16).
Theorem 5’. The absolute convergence of the improper integral (7.16)
yields the convergence of that integral.
5.8
The approximate computation of definite integral
Applying the Newton-Leibnitz formula to evaluate the definite integral,
we have to find the antiderivative of the integrand. A lot of quite a simple
functions, for instance
1
sin x
2
e−x ,
and
x
ln x
don’t have antiderivative among elementary functions. Thus, the NewtonLeibnitz formula is not applicable. In this case we use the approximate
formulas to evaluate the definite integral. One of those approximate formulas
is called trapezoidal rule.
∫b
Let us have an integral f (x)dx for a continuous function f (x) ≥ 0. We
a
divide the interval [a; b] into n subintervals of equal width. So we obtain a
partition
a = x0 , x1 , x2 , . . . xk−1 , xk , . . . , xn = b
where the length of any subdivision [xk−1 ; xk ] is
h=
b−a
n
Hence, xk − xk−1 = h for any k = 1, 2, . . . , n and the dividing points are
x0 = a, x1 = a + h, x2 = a + 2h, ..., xk = a + kh, ..., xn = a + nh = b.
The vertical lines x = xk , k = 1, 2, ..., n − 1 divide the area abBA under
the graph into n areas P QRS. If we substitute the curve between R and S
by the straight line RS, we obtain the trapezoid P QRS, whose parallel sides
P S and QR have the lengths f (xk−1 ) and f (xk ), respectively. The length
of one subdivision h is the height of trapezoid P QRS and the area of this
trapezoid is
f (xk−1 ) + f (xk )
·h
Sk =
2
The sum of the areas of n trapezoids P QRS equals approximately to the
area under the graph abBA. If n is increasing, then the accuracy of this
21
y
S
yk−1
R
yk
y = f (x)
a
P
xk−1
Q
xk
b
x
approximation becomes higher. The area under the graph is the value of the
definite integral. Thus, the definite integral equals approximately to the sum
of the areas of trapezoids P QRS:
∫b
f (x)dx ≈ S1 +S2 +. . .+Sn =
f (x0 ) + f (x1 )
f (x1 ) + f (x2 )
f (n−1 ) + f (xn )
·h+
·h+. . .+
·h
2
2
2
a
Factoring out
∫b
f (x)dx ≈
h
, we have the approximate formula
2
h
(f (x0 ) + 2f (x1 ) + 2f (x2 ) + . . . + 2f (xn−1 ) + f (xn )) (8.17)
2
a
which is called trapezoidal rule. Notice that all the values of the function are
multiplied by 2, except the values at the endpoints y0 = f (a) and yn = f (b).
∫2
Example 12. Compute by trapezoidal rule x2 dx.
0
To compare the result with exact value, we calculate first this definite
2
∫2
8
x3 2
integral by Newton-Leibnitz formula x dx = = = 2, (6).
3 0 3
0
Now we compute this definite integral by trapezoidal formula. First we
divide the interval of integration into four four equal parts [0; 2], that means
2−0
n = 4. The length of one subdivision is h =
= 0, 5 and dividing points
4
are x0 = 0, x1 = 0, 5, x2 = 1, x3 = 1, 5 and x4 = 2.
22
Evaluating the function f (x) = x2 at these points, we have f (x0 ) = 0,
f (x1 ) = 0, 25, f (x2 ) = 1, f (x3 ) = 2, 25 and f (x4 ) = 4. By trapezoidal rule
(8.17)
∫2
x2 dx ≈ 0, 25(0 + 2 · 0, 25 + 2 · 1 + 2 · 2, 25 + 4) = 2, 75
0
Next we compute this integral by trapezoidal rule again, dividing the
interval of integration [0; 2] into eight equal parts. Then the length of one
subdivision is h = 0, 25 and the dividing points are x0 = 0, x1 = 0, 25,
x2 = 0, 5, x3 = 0, 75, x4 = 1, x5 = 1, 25, x6 = 1, 5, x7 = 1, 75 and x8 = 2.
The values of the function f (x) = x2 at these points are f (x0 ) = 0,
f (x1 ) = 0, 0625, f (x2 ) = 0, 25, f (x3 ) = 0, 5625, f (x4 ) = 1, f (x5 ) = 1, 5625,
f (x6 ) = 2, 25, f (x7 ) = 3, 0625 and f (x8 ) = 4. By trapezoidal rule (8.17)
∫2
x2 dx ≈ 0, 125(0+2·0, 0625+2·0, 25+2·0, 5625+2·1+2·1, 5625+2·2, 25+2·3, 0625+4) = 2, 6875
0
23
Download